Ruby on Plain

Plain is a Rails engine that serves as an Artificial Intelligence (AI) assistant for your Rails project. It's not just about organizing your codes or managing your project structure, but about providing deeper, more meaningful context to your work, in real-time. It was proudly developed and presented during the esteemed Rails Hackathon 2023.

This is an example of plain operating in Rauversion.com project

How to use my plugin.

Add this line to your application's Gemfile:

gem "plain-rails"

or development version

gem "plain-rails", github: "chaskiq/plain", branch: "main"

Plain is a Rails engine that serves as an Artificial Intelligence (AI) assistant for your Rails project. It's not just about organizing your codes or managing your project structure, but about providing deeper, more meaningful context to your work, in real-time. It was proudly developed and presented during the esteemed Rails Hackathon 2023. Special Kudos to @claunicole, @silva96

The most salient feature of Plain is its capability to analyze your project on the go. Using AI, it can instantly provide explanations and insights about various aspects of your Rails project, effectively becoming an intelligent companion during your development process.

Beyond just an AI assistant, Plain acts as a comprehensive documentation site. What sets Plain apart is its ability to take markdown files and seamlessly integrate them into the documentation site. No more disjointed files and folders, everything is displayed on a beautifully designed, user-friendly site. This enables developers to add, modify, or access documentation effortlessly and intuitively.

rails plain:install:migrations

Put the following on config/initializers:

config/initializers/plain.rb

Plain.configure do |config|
  config.paths = [
    Rails.root.join("Gemfile"), 
    Rails.root.join("app/models"), 
    Rails.root.join("app/controllers"), 
    Rails.root.join("spec")
  ]
  config.extensions = ["rb", "js", "md", "json", "erb"]
  config.chat_environments = [:development]

  # initialize your vector search
  config.vector_search = Langchain::Vectorsearch::Qdrant.new(
    url: ENV["QDRANT_URL"],
    api_key: ENV["QDRANT_API_KEY"],
    index_name: ENV["QDRANT_INDEX"],
    llm: Langchain::LLM::OpenAI.new(
      api_key: ENV["OPENAI_API_KEY"],
      llm_options: {},
      default_options: {
        chat_completion_model_name: "gpt-3.5-turbo-16k"
      }
    )
  )
end

For the above configuration some environment configuration variables would be required:

QDRANT_URL = 
QDRANT_API_KEY = 
QDRANT_INDEX = 
OPENAI_API_KEY =

You can get a free account on the https://qdrant.tech/. For other vector search please refer to langchainrb repo https://github.com/andreibondarev/langchainrb#using-vector-search-databases-

rails plain:load

In config/routes.rb mount the engine:

mount Plain::Engine => "/plain"

in app/assets/config/manifest.js

add

//= link plain.css

Plain provides a documentation site that can be populated via markdown files with front-matter support, how it works.

Place your markdowns on a docs folder on your project's root. also you can add a main configuration file, for example:

The documentation site will be served at localhost:3000/plain/docs

Put this on /docs/config.yml

name: "Plain docs"
logo: "logo.png"
chat_envs: 
  - "development"
  - "test"
links:
  -
    name: "Start us on github"
    url: "https://github.com/xxx/rxxx
  - 
    name: "X"
    url: "https://x.com/xxx"
footer:
  legend: "xxxx © Copyright 2023 . All rights reserved."
  links:
    - 
      name: "Twitter / X"
      url: "https://twitter.com/xxx"
    - 
      name: "IG"
      url: "https://instagram.com/xxx"
sections:
  - 
    name: "getting_started"
    position: 1
    items:
      - 
        name: "oli"
        path: "aaa"
        description: "hello there"

If you want to have a static (HTML) version of the documentation site you can run the following command:

rails plain:site:compile

This will create a folder on public/static-plain/docs

Right now the script will assume that the site is at plain/docs so it will hit that endpoint, Eventually we will add some options to customize that.

bin/rails plain:tailwind_engine_watch --trace

License: MIT